ANN ARBOR—During a year in which the circulating strains of influenza showed genetic differences from the strains in vaccines, the traditional killed-virus flu shot was found to be effective in preventing influenza in healthy adults. The live attenuated-virus nasal spray vaccine also prevented illnesses but was less effective.

Both outcomes were determined by laboratory confirmation of flu infection, a University of Michigan study found.

Earlier studies had suggested that the nasal spray, sold by MedImmune as FluMist, might offer better protection against drifted viruses that had genetically changed between vaccine formulation and annual influenza activity. The nasal spray, which is based on a live but weakened virus, was 86 percent protective in one study conducted in children during a major drift year. However, FluMist had not previously been studied head-to-head against the shot in adults with laboratory confirmation.

The killed-virus flu shot is usually billed as 70 to 90 percent effective against circulating strains that are well matched to vaccine strains. During the 2004-2005 flu season, a University of Michigan team found that the killed-virus flu shot was 75 percent effective against a moderately-drifted type A virus and two types of B virus. The standard formulation of both the flu shot and the nasal spray vaccine includes two types of A influenza and one B, but in the 2004-05 season, there were two B strains circulating and one type A.

” On the other hand, FluMist was 48 percent effective. These results may only apply to the 2004-05 flu season,” said Dr. Arnold S. Monto, professor of epidemiology. ” In other years the results may be different. We need a more specific understanding of which viral changes matter and which don’t. There are many things about vaccine protectiveness that we still don’t completely understand.”

” It may be that the difference in effectiveness between the shot and the spray can be attributed to poorer protection against type B infections in participants given FluMist,” Monto said, but that’s not clear from this study.

The live attenuated vaccine marketed as FluMist was developed at the University of Michigan by Hunein ” John” Massaab, professor emeritus of epidemiology. MedImmune produces it under a license with the University.

Monto’s research team is conducting a randomized, double-blind, placebo-controlled three-year trial with National Institutes of Health funding in which the two vaccines are being compared head-to-head and against a placebo. Prior to the 2004-05 flu season, they vaccinated 1,247 people aged 18 to 46 in four Michigan communities.

A source of confusion about vaccine effectiveness against drifted viruses may stem from study design. Many vaccine evaluation studies rely on a clinical diagnosis of ” influenza-like illness,” or measures of the immune response to infection in the blood. For the present study, Monto and lead author Suzanne Ohmit, assistant research scientist in epidemiology, took throat swab specimens from participants experiencing flu symptoms and analyzed them, using virus isolation and PCR techniques, to determine if influenza virus was causing the illness.

Monto and Ohmit suspect that the adult participants in their study in 2004-05 had enough prior experience with influenza that the live attenuated virus may have failed to infect their nasal passages and initiate an immune response as it is intended to do. However, both still believe that the nasal spray is very effective in children and may be more effective than the killed vaccine in the young.

” FluMist works very well in children with naive immune systems,” Ohmit said. The FDA approval of FluMist limits its use to people between age 5 and 49 years.

In another paper in the same edition of the New England Journal of Medicine, FluMist was administered to school children and the children were followed to see if they and their families had been protected against flu. The study’s authors conclude that vaccinating children had some protective effect on the children and their families, confirming a notion of ” herd immunity” first tested by Arnold Monto in a 1970 influenza vaccine study.

Without having children in their study, Monto and Ohmit can